AI just hit a new security and deployment phase

AI news just got sharper, and a lot more unsettling: hackers are using models to speed up zero-day exploitation while big tech races to turn AI into the operating layer of your devices. Here are the must-know moves. ^[1][2][4]

AI hacking goes industrial-scale: Google says threat actors are using commercial models to refine and scale attacks, and The Hacker News reports the first known AI-developed zero-day 2FA bypass used for mass exploitation. That’s a real defender wake-up call. ^[1][2]

Android becomes an AI layer: Google says Gemini Intelligence will move across apps, read the screen, build carts, and book reservations before asking for confirmation. In other words, the phone is starting to act less like a tool and more like an agent. ^[4]

Enterprise AI is shifting from pilots to production: Capgemini invested in the OpenAI Deployment Company, and IBM launched managed Red Hat AI Inference on IBM Cloud for real-time production workflows. The big story now is deployment, governance, and scale. ^[5][10]

Regulators are moving too: Colorado watered down its AI law to notification and appeal rights, while Connecticut advanced broader AI rules for chatbots, employment tools, and synthetic content. Meanwhile, TIME says OpenAI, Google, Anthropic, Alibaba, and Mistral are still shaping the market. ^[11][13][18]